Chem. Mater. 1994,6, 461-467
461
Zinc(I1) Oxide Stability in Alkaline Sodium Phosphate Solutions at Elevated Temperatures Stephen E. Ziemniak* and Edward P. Opalka Knolls Atomic Power Laboratory, P.O.Box 1072, Schenectady, New York 12301-1072 Received November 5, 1993. Revised Manuscript Received January 19, 1994"
Zinc oxide (ZnO) is shown to transform into either of two phosphate-containing compounds in relatively dilute alkaline sodium phosphate solutions a t elevated temperatures via ZnO(s) Na+ + HzP04- a NaZnPOr(s) + H2O or 2ZnO(s) + HsPOdaq) is Znz(OH)PO4(s) H20. X-ray diffraction analyses indicate that NaZnP04 possesses an orthorhombic unit cell having lattice parameters a = 8.710 f 0.013, b = 15.175 f 0.010, and c = 8.027 i 0.004 A. The thermodynamic equilibria for these reactions were defined in the system ZnO-NazO-P205-H~0 for Na/P molar ratios between 2.1 and 3. On the basis of observed reaction threshold values for sodium phosphate concentration and temperature, the standard entropy ( S O ) and free energy of formation (AGf") for NaZnPO4 were calculated to be 169.0 J/(mol K) and -1510.6 kJ/mol, respectively; similar values for Znz(OH)P04 (tarbuttite) were 235.9 J/(mol K) and -1604.6 kJ/ mol. Additions of sodium sulfite and sulfate did not alter the above reactions.
+
+
Introduction Recent investigation of the solubility/phase behavior of zinc(I1) oxide in alkaline sodium phosphate solutions1 established a ZnO/NaZnP04 phase boundary in terms of an equilibrium constant for the liquid-phase composition: [Na+l[HzP04-]. According to Thilo and Schulq2 precipitation of a second basic salt of zinc(I1) oxide, Znz(OH)P04 (tarbuttite), is possible under mildly alkaline, hydrothermal conditions. A hydrous form of this mineral is also known to exist: Znz(OH)POp1.5HzO (spencerite). The present work was undertaken to determine (a) the manner in which sulfur oxyanion salts (sulfate and sulfite) affect formation of NaZnPO4 and (b) whether a sodiumfree zinc phosphate compound would form in the system ZnO-NazO-PzO5-H20 for 2 < Na/P < 3. Experimental Section Autoclave Tests. All testa were conducted in a 1-L, goldlined autoclave vessel fitted with a platinum dip tube to permit hot sampling. A sketch of the apparatus is shown in Figure 1. The experimental methodology was the same as that described previously,3 viz., incrementally elevate the temperature of a sodium phosphate solution in contact with a ZnO powder reactant until a phosphate loss is observed. Reagents. Demineralized and deaerated water (resistivity > 1 MQ-cm) was used to prepare all test solutions used in our experiments. All chemicals used [ZnO, NazHPO4 (anhydrous), NasP0~12Hz0,NazSOa, and NazSO4 (anhydrous)] were of analytical or equivalent grade obtained from Fisher Scientific Co. or the GSA Chemical Commodities Agency. Zinc oxide was tested in its as-received condition. Analytical Procedures. The total phosphate concentration and sodium-to-phosphate molar ratio (NaIP) of each sample aliquot were determined by the potentiometric titration procedure described previously.3 The existence of lower phosphate concentrations, however, made it prudent to titrate larger aliquots (up to 50 g of solution) in order to achieve the stated accuracies.
* To whom correspondence should be addressed.
Abstract published in Advance ACS Abstracts, March 1, 1994. (1)Ziemniak, S. E.;Jones, M. E.; Combs, K. E. S. J. Solut. Chem. 1992,24, 1153. (2) Thilo, E.; Schulz, I. 2.Anorg. Allg. Chem. 1951, 265, 14. (3)Ziemniak, S.E.;Opalka, E. P. Chem. Mater. 1993, 5, 690. 0
0897-4756/94/2806-0461$04.50/0
TO HIGH AIR COOLED
THERMAL EXPANSION OVERFLOW
RECORDING THERMOCOU THERMOWEL
GOLD LINER
SAMPLE LINE-
-
-ZINC OXIDE ( 5 B)
1st AUTOCLAVE ( 1 LITER)
2nd AUTOCLAVE (UNHEATED)
Figure 1. Schematic of autoclave arrangement used in zinc oxide transformation study. The above procedure was modified for sulfite determination (in the presence of phosphate), based on the reaction with formaldehyde:
SO:- + CH,O + H,O s (CH,OH)(SO;)
+ OH-
(1)
After titrating to the first phosphate hydrolysis point a t pH = 9.2 f 0.1, sufficient formaldehyde was added to the liquid sample to react completely with the sulfite. The hydroxyl ions thus liberated increased the solution pH, which was then titrated back to the original pH = 9.2 0.1 value. The equivalent amount of titrant used was equal to the amount of sulfite present. The expected accuracy of this procedure is *3%. Sulfate ion concentrations were determined by ion chromatographic analysis with an accuracy of *lo%. Upon completion of each autoclave experiment, the reacted zinc oxide powder was rinsed with deionized water to remove any traces of unreacted sodium phosphate and subjected to visual examination under a stereomicroscope. A portion of the reaction cake was then fractured and placed on a mount for high magnification viewing on a scanning electron microscope. By means of micromanipulators, i.e., needles and tweezers, the reaction product crystals were separated from the remaining mass of unreacted/reacted material. Samples isolated in this manner were subjected to X-ray diffraction (XRD) analyses. Powder
*
0 1994 American Chemical Society
462 Chem. Mater., Vol. 6, No. 4, 1994 6.2
I
I
I
Ziemniak and Opalka
I
Table 1. Zinc(I1) Oxide Phase Transformation Thresholds
I
1.i
~~~~
6.0
5.8
ID
Y
>
E 5.6
I
\e
4.8
4.6 115
indicated threshold final Na/P test DhoSDhaW NalPb temD,K DhosDhateO NdPb loss 1 6.076 2.159 455 3.928 2.775 1.0 2 6.507 2.513 482 3.980 3.748 0.6 3 6.339 2.781 472 3.433 4.260 1.0 4 111.6 2.166 353 85.29 2.540 1.0 5 6.223 2.158 (1) 400 4.349 2.623 1.1 6 5.970 2.458 (2) 443 4.675 2.887 0.9 7 6.781 2.202 (3) 439 5.212 3.300 c 8 5.517 2.452 (4) 455 3.780 3.153 0.9
I
I
135
I
I.
155 175 Temperature, "C
I
I
195
I 215
Figure2. Phosphatevariations caused by heating in the presence of zinc(I1) oxide.
XRD measurements were performed using a Model CN2155D5 Rigaku diffractometer (Bragg-Brentanogeometry) and Cu Ka radiation (A = 1.5417 A). The X-ray tube was operated at 45 kV and 20 ma. Data were taken as a continuous scan from 8 to 92O (26) at a speed of 1°/min. In an attempt to react to completion, test 4 was charged with 2 g of ZnO, rather than 5 g, and operated for 7 days. The reaction product from this test was subjected to additional characterizations by infrared spectroscopy (IFt)and quantitative chemical analyses. The solids IR spectrum was obtained on a PerkinElmer Model 283 spectrometer using the KBr pellet technique. Elemental chemical microanalyseswere performed by Galbraith Laboratories, Inc. (Knoxville, TN).
Results Zinc Oxide Phase Boundary I. Threshold temperatures at which sodium phosphate precipitation was initiated, were determined by reverse extrapolation of phosphate concentration versus temperature plots. At least three temperature points above the threshold levels were included from each run. This process is illustrated in Figure 2. On the basis of changes in the Na/P ratios of the depleted test solutions, compared with initial (baseline) values, indicated Na/P loss ratios were calculated. These values, along with the estimated ZnO transformation temperatures, are summarized in Table 1. No losses of any sulfur oxyanion species (sulfate or sulfite) were observed. The above results indicate that the zinc oxide-sodium phosphate reaction product from all eight tests contained sodium and phosphate in a 1:lmolar ratio. Test 4, which decomposed the entire 0.025 mol charge of ZnO, precipitated approximately 0.026 molof phosphate from solution. This result indicates a reaction product composition Zn/P = 1.
~
a mmol/kg. Na/P ratio for testa 5-8 excludessodium contributed by sulfur oxyanion salts (1)sulfate = 6.933, (2)sulfate = 7.110,(3) sulfite = 8.593,(4)sulfite = 5.333. Na/P loss calculation invalid due to suspected leaching of sodium phosphate from system crevicesfrom a previous run that used extremely high phosphate concentrations.
Although the present experimental program was not designed to verify that threshold conditions for the reverse reaction (i.e., reaction product transformation back to ZnO) were identical to those observed for the ZnO transformation reaction, it is noted that the reversible nature of the ZnO transformation reaction was demonstrated in our previous ZnO/NaZnPOd solubility study.' In that study, reaction product dissolution (increases in sodium phosphate concentration) of a partially transformed ZnO bed and changes in Zn(I1) ion concentration with time were monitored when the sodium phosphate concentration/ temperature conditions were held below the established ZnO reaction threshold conditions. Characterization of Reaction Product I. Individual microcrystals of the reaction product, when viewed under a stereomicroscope,appeared white and translucent. SEM photographs of representative reaction product crystals are provided in Figure 3 at magnifications between 1000 and 10 00OX. These photographs reveal the crystals to have a barrellike appearance. In most instances the "barrels" have a hexagonal cross-section, a hollow center, and an l/d ratio of approximately 211. Crystal sizes (525-pm diameter) are relatively large in comparison to the unreacted zinc oxide (0.5 pm); see Figure 3b. XRD analyses, as performed with monochromatic copper K a X-rays, revealed that the zinc oxide decomposition products from testa 1-8 had identical crystalline lattice configurations; that is, they were the same compound. Only in test 4 were the peak intensities sufficiently strong to provide an X-ray diffraction pattern useful for reference purposes; see Table 2. A search of the JCPDS database4 revealed that although the reaction product possessed some similarities with the a-NaZnPO4 comunder nonhydrothermal pound generated by Kolsi et alms conditions, the interplanar spacings are most consistent with the Na(Zno.aFe0.2)POrcompound generated by Kabalov et al.e via a hydrothermal synthesis. Their reported crystal symmetry is monoclinic with unit cell lattice parameters of a = 8.668 f 0.006, b = 8.125 f 0.004, and c = 15.281 f 0.001 A; B = 90.17O. An indexing of the Table 2 X-ray diffraction pattern by means of the DICVOL computer code7 obtained a satisfactory fit with an orthorhombic unit cell having lattice parameters a = 8.710 f (4) JCPDS Powder Diffraction File, International Centre for Diffraction Data, Swarthmore, PA, 1989; Sets 1-39. (5) Kolsi, A.-W.; Erb, A.; Freundlich, W. C. R. Acad. Sci. Paris, Ser.
C 1976,282,575. (6) Kabalov, Yu.K.; Simonov, M. A.; Mel'nikov, 0. K.; Belov, N. V.
SOU.Phys.-Dokl.
1973,17,835. (7) Louer, D.; Louer, M. J. Appl. Crystallogr. 1972,5, 271.
Chem. Mater., Vol. 6, No.4, 1994 463
Zinc(1Z) Oxide Stability
(a) looOX
Table 2. Indexed Power X-ray Diffraction Pattern of Zinc(I1) Oxide-Sodium Phosphate Reaction Product dd measd d pattern re1 reflection plane indices spacing, A int, 1110 h k I spacing.A 5.M 6 1 0 1 5.903 .... 1 1 4.659 2 4.646 35 4.355 4.360 65 2 0 0 1 4.374 1 3 2 0 4.014 4.008 80 0 2 17 0 1 3.880 3.883 3.712 1 18 2 1 3.710 2 2 3.286 11 1 3.290 4 1 3.192 14 1 3.198 2 3.144 6 0 3.141 3 1 2 3.052 6 3.058 3 2 2.952 2 0 100 2.959 2 1 2.958 3 2 2.860 4 18 0 2.863 2.866 5 0 1 2.629 2 1 20 4 2.629 2.522 3 70 1 1 2.522 2.524 3 2 0 2.413 0 31 6 1 2.414 2.330 2.331 7 2 4 2 2.332 1 2 5 2.280 2 3 0 8 2.280 2.283 3 1 3 2.255 2 1 20 3 2.252 2.184 2 2 3 2.182 9 2.116 2 5 2 27 2.117 2.007 6 0 0 4 2.007 2.007 0 5 3 1.990 4 1 0 1.989 10 1.920 14 2 6 2 1.922 1.834 5 3 3 3 1.834 1.823 2 0 4 9 1.824 1.824 1 3 4 1.823 5 2 3 1.674 4 5 1.616 I4 0 1.651 3 0 4 1.650 5 1.649 4 3 2 1.651 3 3 5 1.650 4 6 0 1.650 1 9 0 1.570 5 1 1 1.513 6 1.571 5 2 0 1.572 4 0 6 1.572 2 0 9 1.530 I4 1 2 9 1.528 1.506 5 23 2 0 1.500 1.507 5 1 3 1.509 4 7 1 1.460 5 1.460 16 0 3 1.458 5 1 4
I
Figure 3. Scanning electron micrographs of zinc oxidesodium phosphate reaction products. Bar length correspondsto 10 and 1 Nm in (a) and (b), respectively.
0.013, b = 15.175 f 0.010.and c = 8.027 f 0.004 A (Le., B = goo). The figure of merit for this fit (Fa1= 4.6) is consistent with an average (absolute) discrepancyin 28 of 0.028O.
Standard quantitative chemical analyses of the test 4 reaction product performed after digestion in nitric acid yielded element Na P (aa PO*) Zn
wt%
11.10 15.88 (49.50) 37.19 total 97.8
atom ratio to P 0.94 1.11
A ZnIP atom ratio slightly in excess of unity, along with approximately 2% unaccounted for material (assumed to be oxygen), is consistent with the presence of &IO% unreacted ZnO in the test 4 reaction cake. The absence
Table 3. Additional Zinc(I1) Oxide Reaction Thresholds indicated baseline f d NaJP test phosphate N d P temp,K phosphate NaJP losa 9 10 11
2.864 2.611 1.885
2.164 2.453 2.150
494 528 481
2.611 2.316 1.485
2.538 2.750 2.735
-1.7 0.1 0.0
of hydroxyl ions was inferred from the absence of an infrared absorption peak in the vicinity of 3500 f 100 cmd, which corresponds to the expected stretching frequency of the 0-H bond. This result is also consistent with the quantitative elemental results. Assembly of the above information gives the ZnO decomposition product stoichiometry as NaIP = ZnIP = 1, or NaZnPOd. Zinc Oxide Phase Boundary 11. Three additional tests were conducted using lower phosphate concentrations; see Table 3. As expected, lower phosphate concentrations led to higher reaction threshold temperatures.
464 Chem. Mater., Vol. 6, No. 4, 1994
Ziemniak and Opalka
Table 4. Comparison of Powder X-ray Diffraction Patterns of Second Zinc(I1) Oxide-Sodium Phosphate Reaction Product and Synthetic Tarbuttite synthetic tarbuttites Composite Pattern of Tests 9-11 d spacing, A re1 int Ill0 d spacing, A re1 int Ill0 6.18 5.37 4.59 3.69 3.266 2.971 2.876 2.771 a.928 2.471 2.416 2.355 2.055 a
100 15 12 45 70 40 45 40 15 50 30 15 30
6.14 5.42 4.58 3.68 3.26 2.97
35 25 20 85 90 35
a a a a
a a a
2.43 2.354 2.054
20 40 25
a
Equilibrium constants for eqs 2 and 3 were defined by
K(2)
~N~Z~PO,~H,O
- a,oaN,+aHgo,-
(4)
In the usual manner, activities (ai) of water and all solid phases were taken to be unity, whereas ionic activity coefficients (rj)were used to relate ionic concentrations [Cjl in molality units to thermodynamic activities. Marshall and Jonesll have shown that an extended DebyeHiickel equation of the form log 7;
Obscured by overlap with ZnO line.
-Z:SdI/(l+
AgI)
(6)
However, the inferred reaction product composition differed from the preceding results in that no sodium losses were observed. The reaction appeared to be self-limiting, undoubtedlydue to the relatively large increases in solution Na/P ratios created by the reaction. That is, the higher Na/P ratios created a test environment where the ZnO phase was once again stable. Characterization of Reaction Product 11. The extremely small amounts of phosphate reacted in this three run series (average = 0.3 mmol of P) made a positive identification by bulk analysis very difficult. XRD analyses of selected regions of the ZnO “cakes” enriched with reaction products, however, confirmed that the reaction product’s major lines were coincident with the major (i.e., 1/10> 10%)lines of synthetic tarbuttite;8 see Table 4. Chemical analyses of individual microcrystals of the reaction product were performed using a JEOL Model JXA-8600 electron microprobe operated a t 15 kV. The results indicated the following atomic ratios: ZnIP = 2.2, / P = 4.5, with estimated uncertainties of f15% (14. /&ly trace levels of sodium were detected, i.e., 45%. Therefore, the above analyses confirm that the second ZnO decomposition product is tarbuttite, Zn2(0H)P04.
with A = 1.5 gives reasonable approximations at temperatures in the range 400-500 K for ionic strengths (I= 0.5cC;Zi2)typical of our tests. In eq 6, Zj is the ionic charge number and S is the limiting Debye-Huckel slope (0.51 at 298 K).12 Combining eqs 2-6 yields
Discussion
This balance was converted to a mathematical equation by introducing the followingequilibrium constants defined in terms of thermodynamic activities (parentheses):
The present experimentsdemonstrate that zinc(I1)oxide was decomposed by relatively dilute sodium phosphate solutions. On the basis of observed reaction product stoichiometries, two ZnO decomposition reaction equilibria were isolated: ZnO(s)
+ Na+ + H2PO;
log Ko, = log Q(2,
+ 2 S d I / ( l + 1.5dI)
(7)
where log Q(2, = -log [Na’l [H2P0;l log K(3, = -log [H3P04]
(8)
An ion electroneutrality balance, which accounted for changes in water,12 phosphate ion,13J4and sulfur oxyanion16J6 dissociation behavior with temperature and ionic strength, was then employed to determine H2P04- and HaPOd(aq) concentrations at the Tables 1and 3 threshold conditions. For the most general case, with sodium phosphate and sodium sulfite present, the balance is [H’I
+ [Na’l
= 3[P043-l + 2[HPO4”1 + [H2P0,1 + 2[S03”l+ [HSOLI + [OH-] (9)
e NaZnP04(s) + H20 (2)
or
2ZnO(s) + H3P04(aq)F? Zn2(OH)P04(s)+ H 2 0 (3) The existence of other ZnO decompositionproducts, such as NaZnz(PO4)(HP04)and NazZn(OH)POr, which have been hydrothermally synthesized in the respective phosphoric acideand trisodium phosphatei0solutions, has been ruled out because of their nonrepresentative Na/P ratios. (8)Milnes, A. R.; Hill, R. J. Neues Jahrb. Mineral., Monatsh. 1977, 1, 25. (9) Kabalov, Yp. K.; Simonov, M. A.; Yakubovich, 0. V.; Belov, N. V. SOU.Phys.-Dokl. 1974,123,627.
with log Kj = bl/T + b2
+ b3 In T + b4T + b d P .
(10) Kabalov, Yu. K.; Simonov,M. A.; Belov, N. V. Sou. Phys.-Dokl.
1972, 17,77.
(11) Marshall, W . L.; Jones,E. V. J . Phys. Chem. 1966, 70,4028. (12) Sweeton, F.H.;Mesmer, R. E.; Baes, C. F. J . Solut. Chem. 1974, 3, 191. (13) Mesmer, R.E.;Baes, C. F. J . Solut. Chem. 1974,3, 307. (14)Treloar, N. C. Central Electricity Research Laboratory Report RD/L/N 270173,1973.
Chem. Mater., Vol. 6, No. 4, 1994 465
Zinc(1l) Oxide Stability
Table 6. Dissociation Behavior of Selected Compounds Ki
bi
b3
b4
b5
ref
94.9734 39.4277 37.7345
-0.097 611 -0.032 540 5 -0.032 208 2 -0.017 459 -0.024 982 -0.025 214 -1.019 062 -0.275 677 -0.057 73
-2 710 870 -810 134 -897 579
Sweeton et al.12 Mesmer and Baed8 Mesmer and Baesla Treloar" Khodakovskii et al.16 Khodakovskii et aL15 Oscarson et al.16 Oscarson et al.16 Oscarson et al.16
b2
KW 31 286.0 -606.522 Ki 17 655.8 -253.198 K2 17 156.9 -246.045 K3 -106.51 7.1340 KI -1424.4 10.192 KII -1629.1 5.733 K I (sulfate) 593 089 -8550.82 KII(sulfate) 119 452 -1974.81 KIP"(sulfate) 9 303 -266.96 log K for dissociation of ion pair, NaS04-. Note that
1343.462 305.682 38.202
-43 593 400 -7 715 600 -462 OOO
inclusion of this species in eq 9 requires modification of eq 11 prior to calculating
reaction threshold values for testa 5 and 6.
Table 6. Solution Chemistry Values Used To Define the ZnO/NaZnPOd and ZnO/Zna(OH)PO, Phase Boundaries test [Na+l, m [H2POrl, m log 61 I S AG,kJ/mol 1 2 3 4 5 6 7
8
1.312 X 1.635 X 1.763 X 2.418 X 2.730 X 2.889 X 3.212 X 2.419 X
le2 1k2
le2 10-'
le2 le2 le2 le2
5.815 X 5.128 X 2.851 X 2.188 X 1.258 X 2.179 X 1.124 X 1.391 X
10-4 10-4 10-4 lo" 10-4 10-4 10-4 10-4
5.1173 5.0767 5.2992 5.2765 5.4642 5.2010 5.4424 5.4730
0.01861 0.02243 0.02369 0.3711 0.04034 0.04177 0.04314 0.03123
0.7640 0.8374 0.8057 0.5798 0.6424 0.7355 0.7265 0.7640
-46.08 -48.73 -49.60 -38.71 -43.03 -46.06 -47.67 -49.53
[Hap041 9 10 11
6.198 X 10-3 6.405 X 10-3 4.053 X 10-3
8.795 X leQ 1.320 X 1o-S 7.052 X leg
8.260 X 10-4 9.204 X 10-4 5.868 X 10-4
The above parameters are tabulated in Table 5. Upon substitution of the eq 10 dissociation constants into eq 9 and applying the known levels of total dissolved sulfite, [mJ, and phosphate, [m,] (and its associated sodiumto-phosphate molar ratio, y ) , the neutrality balance reduced to an algebraic equation in terms of the unknown [H+], i.e., x : x
+ (Y-3)mP =
X
(11)
Equation 11 represents a seventh-order polynomial equation in terms of the unknown, x , and was solved by a Newton-Raphson iteration procedure. The calculated reaction threshold values for [H+l, along with those for [H2P04-] and [HaPOl(aq)l, are given in Table 6. It is noted that measured and calculated pH values (at room temperature) for all baseline test solutions agreed to within the reproducibility of the measurements (f0.05). All AG values reported in the final column of Table 6 were determined from the expression
Least-squares fits of the Table 6 results (see Figures 4 and 5) yield (15) Khodakovskii,I. L.;Ryzhenko, B. N.; Naumaov, G.B. Geochem. Int. 1968, 5, 1200. (16) Oscarson, J. L.;Izatt, R. M.;Brown, P. R.; Pawlak, 2.;Gillespie, S.E.; Christensen, J. J. J. Solut. Chem. 1988, 17, 841. (17) Criss, C. M.; Cobble, J. W. J . Am. Chem. SOC.1964, 86, 5390. (18) Abraham, M. H.; Marcus, Y. J. Chem. Soc., Faraday Trans. 1 1986,82, 3255.
0.00824 0.00810 0.00535
ZnO/NaZnP04: AG (kJ/mol) = -8522
-76.18 -79.64 -75.06
- (86.06 f 12.28)T
(13)
f 1698 - (98.44 f 3.39)T
(14)
f 5395
ZnO/Zn2(OH)P04: AG (kJ/mol) = -27640
Eguation 13 provides free energy changes for eq 2 with an estimated standard deviation of f1.15 kJ/mol over the temperature range 350-490 K. An independent estimate of the eq 2 equilibrium, based on our previous ZnO solubility study,' is plotted for comparison purposes in Figure 4. Given the precision of the latter estimate ( l u = f3.0 kJ/mol), the observed 2 kJ/mol difference between the two correlations is not significant. Combination of the above results with previously established thermodynamic properties yields the summary presented in Table 7. The observed susceptibility of ZnO to hydrothermal degradation in alkaline sodium phosphate solutions has led to the precipitation of sodium-zinc(I1) phosphate phases that are unique with respect to neighboring divalent transition metal ions: Na+ and OH- have not precipitated concurrently. That is, NanZn(0H)POcwas not observed. Although the latter compound has been hydrothermally formed in trisodium phosphate and sodium sulfate sohtions,lo the present results confirm that NaZnPO4 is the preferred zinc(I1) oxide decomposition product at lower solution alkalinities. It is noteworthy that the orthorhombic structure of NaZnPO4 is virtually identical to that possessed by (19) Wagman, D. D.;Evhs, W. H.; Parker, V. B.; Schumm, R. H.; Halow, I.; Bailey; S. M.; Churney, K. L.; Nuttall, R. L. J. Phys. Chem. Ref. Data 1982, 11, Suppl. 2, (20) Khodakovskii, I. L.; Elkin, A. E. Geokhimiya 1975, 10, 1490. (21) Kubaschewski, 0.; Alcock, C. B. Metallurgical Thermochemistry; Pergamon Press: Oxford, 1983. (22) Larson,J. W.; Zeeb, K. G.; Hepler, L. G. Can. J. Chem. 1982,60, 2141. (23) Vieillard, P.; Tardy, Y. In Phosphate Minerals; Nriagu, J. O., Moore, P. B., Eds.; Springer-Verlag: Berlin, 1984, Chapter 4. (24) Magalhaes, M. C. F.; deJesus, J. P.; Williams, P. A. Mineral. Mag. 1986, 50, 33.
Ziemniak and Opalka
466 Chem. Mater., Vol. 6, No. 4, 1994
Table 7. Thermochemical Parameters for Selected Species in the ZnO-Na20-PaOrHaO System at 298.15 K Cpo (298), So(298), W0(298), AGfo(298), species J mol-' K-' J mol-' K-' kJ mol-' kJ mol-' ref H+(aqP -71 f 1 -22.2 0 0 17,18 Na+(aq) -24.6 36.8 -240.12 -261.91 19 Znz+(aq) -164 -154.8 f 1.3 -153.64 f 0.42 -147.23 18,20 Hz(g) 28.83 130.58 f 0.08 0 0 21 0 0 21 Oz(g) 29.37 205.02 f 0.4 HzO 75.31 69.96 f 0.08 -285.85 f 0.04 -237.19 f 0.04 21 -153.6 -1277.4 -1018.8 19,22 P04Vaq) -283 f 20 -112 f 4 -32.6 -1305.5 -1089.7 14,22 HPOP(aq) 72.4 -1308.8 -1130.8 13,22 HzPOr(aq) 37 f 4 HsPOr(aq) 94 f 4 119.8 -1300.2 -1143.0 13,22 0 0 19 NaW 28.24 51.21 P(S) 23.84 41.09 0 0 19 ZnW 23.40 41.63 0.21 0 0 21 €-Zn(OH)z(s) 72.4 76.99 f 0.21 -645.47 -555.93 f 0.21 20 -350.83 f 0.21 -320.91 f W26. 21 ZnO(s) 40.25 43.64 & 0.42 cu-Zns(PO4)r4HzO(s) 372.7 -4102.0 -3628.9 23 Zna(POdz.2HzO(s) 243.0 -3516.3 -3143.5 23 ZndPO4)z~HzO(s) 184.3 -3211.7 -2890.9 23 Zns(POdz(s) 134.3 -2899.6 -2633.4 23 24 Zn~(OH)P04.1.5HzO(s) -1982.4 f 3.1 Znz(OH)PO4(s) 235.gb -1743.6 -1604.6 this work -1632.1 f 4.0 24 1,this work NaZnPO&) 150.3b 169.0 -1622.4 -1510.6
*
0 Tabulated CPo and So values differ from the usual convention of zero due to conversion to an absolute scale of ionic properties. b Mean values over temperature range of this study.
Table 8. Estimated Reaction Thresholds for Zinc(I1) Oxide Decomposition in the ZnO-PtOrHaO Ternary Oxide System at 298.15 K decomposition reaction 2ZnO(s) + H*Ol(aq) + l/zHzO Zn~(OH)PO~1.5HzO(s) 2ZnO(s) + HsPOl(aq) Znz(OH)PO4(s)+ HzO 3ZnO(s) + 2H$Ol(aq) + HzO F? (Y-Z~~(PO~)Z.~HZO(S) 3ZnO(s) + 2H$Ol(aq) Zn3(P04)2*2HzO+ HzO 3ZnO(s) + 2HsPOl(aq) F? Zn3(PO*)z.HzO + 2Hz0
AGO, kJ/mol
* *
beryllonite (NaBeP04):26a = 8.16,b = 14.08,and c = 7.79 A; the lower lattice constants being caused by the smaller size of the Be(I1) ion relative to the Zn(I1) ion. On the other hand, the structural similarity between NaZnPO4 and the Na(Zno.$eo.2)PO4 compound reported by Kabalov et al.6 indicates that zinc(I1) oxide decomposition in the presence of solubilized Fe(I1) ions may lead to crystallization of an "impure" sodium zinc phosphate precipitate that represents a solid solution of Zn(I1) and Fe(I1) ions - Na(Znl-,Fe,)PO4. The incorporation of Fe(II),however, may cause slight lattice distortions which lead to a monoclinic cell, Le., 90 90.2'. To more fully interpret the significance of formation of a sodium-less ZnO decomposition product in alkaline sodium phosphate solutions having Na/P 1 2.0, a comparison of equilibrium constants associated with the formation of some zinc phosphate minerals expected to exist in the ZnO-PzOa-HzO ternary oxide system is given in Table 8. These estimates were derived from recent room temperature solubility studies for a-hopeite,26tarbuttite,24 and spencerite.2' On the basis of the Table 7 compilation,AGO estimates (henceequilibrium constants) were obtained for potential ZnO decomposition reactions. As expected, a Zna(PO& solid having four waters of hydration is the stable zinc phosphate phase at room temperature, and progressive dehydration to two, one, and zero waters of hydration may be expected as temperature increases. On the other hand, an anhydrous Zn2(0H)P04 solid is expected to be thermodynamically stable with respect to the hydrated form (spencerite) at room tem-
-
(25) Golovastikov, N. I. Sou. Phys.-Crystallogr. 1962,6, 733. (26) Nriagu, J. 0. Geochim. Cosmochim. Acta 1973, 37, 2357.
log tH~POdas)l,moVkg -13.831 -14.791 -12.518 -11.533 -10.200
-78.94 -84.42 -142.89 -131.65 -116.43
-
. z --
Phosphate Only 0 Sulfate Added 0 Sulfite Added
-40
-
b -
8b.45 f i
-50
8
-
-
-
-
-55
1
!
I
perat~re.~'The calculated threshold concentrations of HaPOr(aq) required to initiate these reactions are compared in Table 8. This comparison reveals that the tarbuttite formation reaction requires the lowest concentrations of HsPOr(aq) in order to procede. Since the initial solutions for tests 9-11 had room-temperature log [HaPOs(aq)l values that ranged between -14.4and -15.5 and were higher than in the first eight runs, it is seen that temperature increases were sufficient to cause the ZnO/ Zn2(0H)P04 phase boundary to be crossed. An inconsistency exists with regard to the ZnO/Znz(OH)P04 phase boundary, since the present three hightemperature points do not extrapolate linearly to the implied room temperature estimate2' (see Figure 5). It is noted that the initial (i.e., room temperature) HsPO'(aq)
'
Chem. Mater., Vol. 6, No. 4, 1994 467
Zinc(ll) Oxide Stability -60 -65
an upper limit of AGO = -76.4 kJ/mol to be derived for eq 3 using the results of Magalhaes et aLZ4This estimate differs by -19 kJ/mol from one based on extrapolation of the eq 3 equilibrium via eq 14: AGO = -57.0 kJ/mol. Although some of the observed discrepancy may be due to a heat capacity effect (i.e., curvature in the AG versus T correlation), the absence of heat capacity data for Znz(OH)PO4does not allow this contribution to be estimated. Therefore, the presently extrapolated thermochemical parameters for tarbuttite are considered tentative, pending completion of a ZnO phase stability study in sodium phosphate solutions having Na/P I2.
f -75
-85 280
I 320
360
1
1
400 440 Temperature, K
1
I
400
520
I
560
Figure 5. Free energy changes determined for zincite transformation to dizinc hydroxyphosphate (tarbuttite) in alkaline sodium phosphate solutions.
concentration used in test 11was greater than the implied HaPOr(aq) threshold concentration for eq 3, based on the results of Magalhaeset aLZ4Therefore, the true AGO value for eq 3 must be greater than -81.98 kJ/mol. Considering the reported uncertainty in AGfO for tarbuttite formation (f4.0 kJ/mol per Table 7) and invoking a 2a limit, allows
Acknowledgment. We are indebted to the following individuals who contributed professional assistance: P. C. Sander for X-ray diffraction analyses; Dr. J. J. Cheung for infrared spectroscopy; G. M. Neugebauer for scanning electron microscopy; and V. P. Nordstrom for electron microprobe quantitative chemical analyses. The Knolls Atomic Power Laboratory is operated for the US.Department of Energy by KAPL, Inc., a wholly owned subsidiary of the Martin Marietta Corp., under Contract DE-AC12-76-SN00052.
